Large Language Models in Multi Criteria Decision Making: A Systematic Review, Taxonomy, and Future Research Agenda

Abstract

This study is a systematic review, bibliometric analysis and taxonomy building of Large Language Models (LLM) and Multi-Criteria Decision-Making (MCDM) integration. A total of 63 publications from the Scopus database were analysed using PRISMA guidelines and VOSviewer tools. The findings show this field is highly nascent, with a rapid rise since 2025. Bibliometric analysis shows that this field is widely practiced globally, with 37 countries involved and China and India contributing most publications. This study introduces a new classification system based on the roles of LLM, the integration stages and the application areas. Comparative studies show LLM-enhanced MCDM improves automation, scalability, and unstructured data processing, but remains prone to biases, lack of interpretability and sensitivity to prompts. The research concludes with gaps and opportunities to advance this field towards robust, explainable and hybrid intelligent decision-making systems.

Ähnliche Arbeiten